clp-research / slurk-bots Goto Github PK

Spent a couple of afternoons off and on trying to build an image for the dito bot before I figured out that that the Dockerfile was written to be invoked from the directory above.

That is, I was in the dito/ directory and calling podman build . -t slurk/dito and getting the file not found error that pops up sometimes with Docker COPY commands. When I edited the Dockerfile relative to this directory (removing dito/ in line 6 and replacing dito with . in line 9) it started working and only then did I realise my error.

Might be good to add instructions in the slurk-bots/README.md or the README for each individual bot for how folks are supposed to build the images.

(I also tested with docker instead of podman+buildah, of course, but had the same silly mistake)

When not providing data for a request, no content type is required. However, the API returns 415 when omitting the content type in any case.

What is the current status of the development branch? There were no contributions in the past year.

All bot docker images are named slurk/<name>-bot
All bot classes are named <Name>Bot
I like that those follow a pattern.
However, our folders are named either <name>bot or only <name>

If we intend to use a template for bots, it would be positioned in the parent directory of all bot scripts #33 . The only two ways, I know of, how to import from there, are to manually manipulate syspath:

ROOT = os.path.dirname(os.path.abspath(__file__))
sys.path.append(ROOT)

Or to execute the bot scripts as modules from the root directory and use absolute imports from there (as seen in #33):

python -m echo

I prefer the later, but at least on WSL (Windows 10, Ubuntu 20.04 LTS) + conda, the python interpreter chooses the standard library module math over a folder in the current working directory with the same name.
So the name math for the MathBot can't be used for this approach.